Hi all,
Today, a machine with vanilla 2.4.20 hung with this in syslog:
kernel BUG at highmem.c:169!
invalid operand: 0000
The complete log can be found at
<http://www.cs.tau.ac.il/~didi/kern-bug-1-orig>, with a
hopefully well-decoded ksymoopsed output at
<http://www.cs.tau.ac.il/~didi/kern-bug-1-ksymed>.
When this happened, the machine still replied to ping, but not
to other things I tried (but I did not try much). The console
did not respond to anything, including Alt-SysRq-B (which is
verified to normally work).
I ran memtest86 for about an hour, no problems, and rebooted.
It's a Dual PIII-1000 with 4GB RAM, CONFIG_HIGHMEM4G=y (full
config here <http://www.cs.tau.ac.il/~didi/config-2.4.20-net1>),
with root (and everything else) on NFS. This might (or might not)
be related to a problem solved in 2.4.21-pre5, noticed as
"xdr nfs highmem deadlock fix" in the changelog.
google found no other BUG in this line (but did find in others).
Is this a known bug? A solved (in which version) bug?
Should I do anything else to make it not happen?
Tell me if you need more info.
I am not on lkml, so please CC me.
Thanks a lot!
Didi
>>>>> " " == Yedidyah Bar-David <[email protected]> writes:
> Hi all, Today, a machine with vanilla 2.4.20 hung with this in
> syslog:
> kernel BUG at highmem.c:169! invalid operand: 0000
Known problem (see the L-k archives). The following patch fixes it...
Cheers,
Trond
--- linux-2.4.21-pre4/fs/nfs/symlink.c.orig 2002-08-14 05:59:37.000000000 -0700
+++ linux-2.4.21-pre4/fs/nfs/symlink.c 2003-02-25 20:42:39.000000000 -0800
@@ -46,7 +46,6 @@
error:
SetPageError(page);
- kunmap(page);
UnlockPage(page);
return -EIO;
}